At ASCO 2014, I provided the commentary after a key presentation demonstrated that the second generation EGFR tyrosine kinase inhibitor (TKI) Gilotrif (or Giotrif in some parts of the world) (afatinib) as first line therapy compared to standard chemotherapy in EGFR mutation-positive patients gave a significant benefit in overall survival (OS) that hadn't been seen when other first generation EGFR TKIs, namely Iressa (gefitinib) or Tarceva (erlotinib) were compared to chemotherapy in EGFR mutation-positive patients. The question was whether this difference meant that Gilotrif is a significantly better EGFR TKI than Iressa or Tarceva. I noted that while these results were provocative, trials with Iressa and Tarceva were far smaller and in most cases were stopped early due to early results showing that the EGFR TKI was clearly superior in short term measures like progression-free survival (PFS) and response rate.  Accordingly, we couldn't say anything definitive about the efficacy of one EGFR TKI vs. another by making inferences of each compared with an increasingly irrelevant comparator. If trial after trial showed that the EGFR TKI was clearly better than chemotherapy, we can't draw meaningful conclusions of one being better based on how much stronger one looked than another against an inferior option. Instead,  the only reliable way to compare two EGFR TKIs would be to directly compare two EGFR TKIs in a randomized trial with the same eligible population.  In fact, such a trial, called LUX-Lung-7, had already been not only conceived but enrolled, randomizing EGFR mutation positive patients in Asia, Europe, Canada, and Australia to either Iressa or Gilotrif. We just needed to see how it turned out.

Fast forward 1.5 years, and we now have the early results of LUX-Lung-7, and they show that these EGFR TKIs are not completely interchangeable. As further background, the study randomized 319 patients to either Iressa at 150 mg daily or Gilotrif 40 mg daily, and at the time of the report, the median follow-up was just over 27 months.  The trial population was about 55% Asian and 45% non-Asian,and 2/3 of the patients were never-smokers. All patients had one of the two major activating EGFR mutations: 58% had the exon 19 deletion, and 42% had the L858R substitution.

The main findings were that while overall survival is still immature (too many patients are alive to call a winner -- a good problem), Gilotrif appeared superior in efficacy as measured by PFS, RR, and "time to treatment failure", which is similar to PFS but also includes factors of patients being allowed to stay on treatment due to the perception that they were still benefiting from the treatment, or coming off due to significant side effects. And while one of the leading concerns about Gilotrif has been that its challenging side effect profile counterbalanced any potential minor difference favoring its efficacy, the head to head trial shows less of a difference than most of us would have expected.

Getting to specifics, the most relevant variable reported here, I would say, is PFS, with scans reviewed not only by the investigator caring for the patient but an independent radiologist not involved with the case. Interestingly, there was no difference in median PFS -- the time when half of the patients have progressed -- for the recipients of Gilotrif vs. Iressa, but after the point where half of the patients had progressed, the patients who hadn't progressed did better with Gilotrif, so that when we look at specific time points beyond a year, significantly more Gilotrif patients remain on their first line therapy (27% vs. 15% at 18 months, 18% vs. 8% at 24 months).

Looking at some of the other efficacy endpoints, median time to treatment failure also favored Gilotrif (13.7 vs. 11.5 months), as did objective response rate (70 vs. 56%) and duration of response (10.1 vs. 8.4 months). Importantly, these trends were seen across every subgroup, regardless of patient sex, smoking status, age, or specific EGFR mutation. As I noted, results for OS remains immature but shows a trend (non-significant) toward being superior for Gilotrif at this time.

 As expected, far more patients receiving Gilotrif required a dose reduction (41.9% vs. 1.9%), but the rate of drug-related serious AEs was less different between the 2 TKIs (Gilotrif - 10.6%, Iressa 4.4%) than I would have expected, and there was absolutely no difference in the rate of discontinuation of treatment due to side effects (6.3% for both). More patients had moderate to severe diarrhea and rash with Gilotrif (12% and 9%, respectively), but 7.5% of patients on Iressa experienced moderate to severe elevation in liver function tests, which wasn't seen with Gilotrif. 

Putting all of it together, what does this mean? Is Gilotrif the clear EGFR TKI of choice? I wouldn't go that far, but the trial does show that a subset of patients with an EGFR mutation do better with Gilotrif than with Iressa, at least in terms of the relatively short-term outcome of response rate and PFS.  At this point, there is only a non-significant trend toward better survival with Gilotrif, and I expect that in a world with more effective drugs for acquired resistance like the third generation EGFR TKI Tagrisso (osimertinib), at least for the 50-60% of EGFR mutation-positive patients who develop a T790M mutation, the relatively modest difference in efficacy of first line therapy will become blunted and is very likely to emerge as significant. How important is PFS? I've argued that it depends on the magnitude of the difference, and this one isn't huge. On the other hand, I also think the threshold of how big a difference of PFS without survival benefit you need depends on the situation: specifically, if you're thinking about adding a very expensive or toxic treatment, it should be huge, but if you're just deciding between two very comparable treatments, why wouldn't you select the one that gives you significantly better short-term results and at least trends towards better long term results?  

One more important finding was that despite some earlier work suggesting that Gilotrif may not be a strong choice for patients with an L858R substitution, the results were better for Gilotrif in the patients with both common activating mutations and no trend of different results (though the patients with an exon 19 deletion had a median PFS 1-2 months longer than the L858R substitution population with either EGFR TKI).  We can conclude that Gilotrif isn't a treatment only for patients with an exon 19 deletion. 

But the comparison in LUX-Lung-7 is of Gilotrif vs. Iressa. in the US, the overwhelming leading first line treatment for EGFR mutation-positive NSCLC is Tarceva. Can we assume that Gilotrif is definitely better than Tarceva? I'd say no.The very poor data we have to compare them is mixed but has at least suggested to me that Tarceva is perhaps marginally superior to Iressa, so I wouldn't presume that Tarceva will fall closer in efficacy results to Iressa than Gilotrif. That said, Gilotrif is the only EGFR TKI we have that has shown superior efficacy directly compared to another, an important finding that I believe should factor into our decisions about which EGFR TKI to recommend to our EGFR mutation-positive patients.

What do you think? If you were prescribed an EGFR TKI other than Gilotrif, is this enough to request a change? Or does the lack of a longer-term demonstrated benefit in overall survival suggest it's not a very clinically significant difference?